Reliability of Meat, Fish, Dairy, and Egg Intake Over a 33-year Interval in Adventist Health Study 2

Nutr Cancer. 2014;66(8):1315-21. doi: 10.1080/01635581.2014.956249. Epub 2014 Oct 8.

Abstract

We studied Adventist Health Study 2 (AHS-2) cohort members to determine the reliability of long-term recall of adult dietary intake that occurred 33 years ago. Establishing the reliability of these measures supports studies of how dietary exposure across the life course affects risk of cancer and other noncommunicable disease outcomes. Among 1816 AHS-2 cohort members, we conducted a statistical comparison of long-term recall of meat, fish, dairy, and eggs at AHS-2 baseline with their report of current diet 33 years before AHS-2 baseline at an age of 30-60 years. Major findings are as follows: 1) a high correlation for frequency of red meat (R = 0.71), poultry (R = 0.67), and fish (R = 0.60); lower correlations for dairy (R = 0.19) and eggs (R = 0.28); 2) good concordance for dichotomous measures of red meat [sensitivity: 0.70; specificity: 0.92; positive predictive value (PPV): 0.91], poultry (sensitivity: 0.76; specificity: 0.87; PPV: 0.83), fish (sensitivity: 0.61; specificity: 0.93; PPV: 0.89), dairy (sensitivity: 0.95; specificity: 0.57; PPV: 0.99), and eggs (sensitivity: 0.95; specificity: 0.41; PPV: 0.96); negative predictive value for dairy and eggs was poor. Among older AHS-2 cohort members, we found good reliability of recall of red meat, poultry, and fish intake that occurred 33 years earlier.

Publication types

  • Research Support, N.I.H., Extramural

MeSH terms

  • Adult
  • Animals
  • Body Mass Index
  • Cohort Studies
  • Dairy Products*
  • Eggs*
  • Feeding Behavior*
  • Female
  • Fishes
  • Follow-Up Studies
  • Humans
  • Male
  • Meat*
  • Mental Recall
  • Middle Aged
  • Neoplasms / epidemiology
  • Poultry
  • Reproducibility of Results
  • Risk Factors
  • Surveys and Questionnaires